A plant in Brazil does not have a recruitment project; it has replacement. Every week a shift loses someone, and the aggregate never looks large enough in any single week to warrant attention while consuming an entire people function and leaving lines short-staffed.

Time-to-cover, not cost-per-hire

The cost of an open industrial role sits in the operation: shifts running short, overtime, adjusted line rates, and the strain on those who stayed — which tends to produce the next resignation. Optimising cost-per-hire while roles stay open for three weeks saves in the wrong place.

Certification is planning, not verification

Continuous replacement hits the same constraint every month: candidates without valid NR-10, NR-33, NR-35 or forklift certification cannot enter the area. At steady state this stops being a check and becomes planning — part of every intake will need training.

An embedded team keeps this mapped: who holds what, what expires when, and which refresher cohort is scheduled. That reduces time-to-cover more than any improvement to job advertising.

The industrial cluster is a closed market

In an industrial cluster, a small number of large employers compete for the same profiles, know each other and know each other's pay levels. Hiring directly from a neighbour carries relationship consequences, and a plant's reputation as an employer circulates quickly between shifts.

In that kind of market, a regional candidate base built over the year is worth more than advertising — and it is exactly what a per-role engagement never accumulates.

Employment runs under CLT with premiums for night work and for hazardous or unhealthy conditions, so rota design changes cost directly. Follow-up continues for 90 days after each admission and reports departures by cause and by shift — the per-shift reading is what separates a selection problem from a leadership problem on a specific rota.